This proposal is designed to continue our efforts to improve the identification of infants at risk for cognitive deficits and to better understand the nature and extent of the newly uncovered infant-child continuities in intellectual functioning. These recent demonstrations of predictive validity from infancy into later childhood are of enormous theoretical significance, indicating that cognitive abilities can be assessed in infancy, that individual differences in these abilities can be detected, and that some of these differences are related to later cognition. The primary aim of the present proposal is to begin to elucidate the basis for this continuity. Our strategy for doing this will be to examine measures of four cognitive processes that are thought to be implicated in the cross-age relations uncovered to date. These are the following: memory, speed of information processing, attention, and representational competence. Differences in any or all of these could underlie individual differences in infant information processing as well as the relations that have been observed between infant measures and later IQ. These aims will be pursued in a longitudinal study of fullterms and high-rink preterms (<1500g) in which infants are first seen at 7 months and subsequently at 12, 24, and 36 months. The specific aims are to (a) assess the independence, specificity, and stability of performance on measures designed to tap these four areas, (b)examine the extent to which infants' ability in each area is related to later IQ, (c) identity the processes that underlie the relation between visual recognition memory (one of the principal infant measures) and later IQ, (d) more precisely and fully specify the areas of cognitive function in which preterms show deficits, and (e) examine the role of behavioral, social, environmental, and medical factors in mediating or moderating infant-child continuities. Understanding the nature of early cognition, its developmental course, and the qualities of mind that form the basis for longitudinal relations is of fundamental importance for understanding the ontogeny of later deficits, for developing early diagnostic instruments, and for the rational design of intervention programs.